An Integrative Model of Cellular States, Plasticity, and Genetics for Glioblastoma

Abstract

Diverse genetic, epigenetic, and developmental programs drive glioblastoma, an incurable and poorly understood tumor, but their precise characterization remains challenging. Here, we use an integrative approach spanning single-cell RNA-sequencing of 28 tumors, bulk genetic and expression analysis of 401 specimens from the The Cancer Genome Atlas (TCGA), functional approaches, and single-cell lineage tracing to derive a unified model of cellular states and genetic diversity in glioblastoma. We find that malignant cells in glioblastoma exist in four main cellular states that recapitulate distinct neural cell types, are influenced by the tumor microenvironment, and exhibit plasticity. The relative frequency of cells in each state varies between glioblastoma samples and is influenced by copy number amplifications of the CDK4, EGFR, and PDGFRA loci and by mutations in the NF1 locus, which each favor a defined state. Our work provides a blueprint for glioblastoma, integrating the malignant cell programs, their plasticity, and their modulation by genetic drivers.

A molecularly distinct subset of glioblastoma requires serum‐containing media to establish sustainable bona fide glioblastoma stem cell cultures

“…The ESCs formed a discrete group and were also in this analysis the closest related to mesenchymal SFC samples (Figure b). In addition, gene expression heatmaps of the recent MES1 (Figure S3a) and MES2 (Figure S3b) meta modules (Neftel et al, ) supported an enrichment of mesenchymal subtype genes in the ESC samples.…”

“…From differential gene expression analysis of ESC_to_SFC and SFC samples, we selected genes showing a log fold change bigger than 1 and a false discovery rate (FDR) value less than 0.05 for pathway analysis using the Reactome data analysis tool (http://reactome.org) with default settings (Fabregat et al, ). Heatmaps based on previously described MES1 and MES2 meta modules (Neftel et al, ) were constructed using R‐library gplots.…”

“…In GBM, this was recently supported through experimental analyses and mathematical modeling, which showed that patient‐derived GBM cell cultures contained multiple subpopulations that were stem cell‐like, tumorigenic, and reversibly adaptable to environmental cues (Dirkse et al, ). An extensive single‐cell RNA sequencing‐based analysis of GBM tumors corroborated the plasticity model, where four main cell states were identified to drive GBM growth and transitions between states dictated by the specific genetic alterations and microenvironment in each tumor (Neftel et al, ).…”

A Modular Master Regulator Landscape Determines the Impact of Genetic Alterations on the Transcriptional Identity of Cancer Cells

“…However, most mutations are not as restrictive. In glioblastoma, for instance, there is only weak association between mutational and transcriptional states (Neftel et al, 2019). Indeed, EGFR mutations, while more frequently associated with a proneural identity, are also detected in mesenchymal cells.…”

Cancer avatars derived from genetically engineered pluripotent stem cells allow for longitudinal assessment of tumor development

“…Additionally, while human astrocytes engineered with combinations of human TERT and H-Ras expression and inhibition of the TP53 pathway either by SV40 T/t-Ag or by HPV E6 and E7 generate gliomas with high-grade histology 5,6 , how well these models recapitulate the full spectrum of glioma pathobiology, especially in terms of GBM heterogeneity, has not been well defined. In contrast, patient derived xenografts (PDX) have been useful to study inter-and intra-tumoral heterogeneity 7,8 and sensitivity to pathway-specific therapies 9 , however they do not allow for experimental standardization or afford analysis of the effects of molecular subtype mutations on tumor evolution.…”